CARM1/PRMT4 facilitates XPF-ERCC1 heterodimer assembly and maintains nucleotide excision repair activity

Abstract

The structure-specific endonuclease, XPF-ERCC1, plays a central role in DNA damage repair. This nuclease is known to be important for nucleotide excision repair, interstrand crosslink repair, and DNA double-strand repair. We found that the arginine methyltransferase, CARM1/PRMT4, is essential for XPF stabilization and maintenance of intracellular protein levels. Loss of CARM1 results in a decrease in XPF protein levels and a concomitant decrease in ERCC1 protein. A similar destabilization of XPF protein was observed in cells expressing a mutant in which XPF arginine 568 was replaced by lysine. Loss of CARM1 impaired XPF-ERCC1 accumulation at the site of damage and delayed removal of cyclobutane pyrimidine dimers by UV. As a result, CARM1-deficient cells showed increased UV sensitivity. Our results provide insight into the importance of CARM1 not only in the mechanism of XPF-ERCC1 complex stabilization but also in the maintenance of genome stability.

Graphical Abstract

Figure 1.

CARM1 deficiency suppresses XPF–ERCC1 accumulation at DNA damage sites and delays CPD removal. (A) CARM1 KO inhibited CPD removal ability. The defect of CPD removal was rescued by exogenous CARM1-WT but not by inactivated CARM1-AAA. Data are shown as the mean ± SD of biological three independent experiments (multiple comparisons: two-way ANOVA, **P < .01, ****P < .0001). (B) CARM1 KO and KD cells showed increased sensitivity to UV light. Biological three independent experiments were performed. Error bars represent SD *P < .05, **P < .01, and ***P < .001 (Student’s t-test). (C) Accumulation of XPC at local UV-exposed sites was not different between siCt and siCARM1, but XPF and ERCC1 accumulations were inhibited by siCARM1. Representative images of immunostaining for XPC and XPF and phosphor HBO1 and ERCC1 in local UV-irradiated HeLa cells. HeLa cells were irradiated with 50 J m⁻² UV through an 8-μm pore membrane and fixed after 30 min of incubation. The mean intensity of the irradiated area was divided by the intensity of the non-irradiated area of the same nucleus and expressed as a relative intensity increase from the baseline (%). Scale bars, 5 μm. Data are shown as the mean (unpaired t-test). (D) CARM1 interacts with XPF in HEK293 cells. (E) XPF interacts with CARM1 in DOX-induced XPF stable expression cells and 293 cells. (F) XPF WT-Myc and GST-CARM1 interact in a GST pull-down assay. When XPF WT-Myc overexpressed in HEK293 cells was pulled down using GST-CARM1 expressed and purified in E. coli, XPF WT-Myc specifically co-precipitated. *An unknown E. coli protein that is co-purified when GST-CARM1 is purified from E. coli. (G) CARM1 KD cells showed decreased total protein content of XPF and decreased chromatin localization. XPF amounts were corrected for ORC2. The amount of XPF in each fraction is expressed as relative amount (%) to the siCt soluble fraction without UV. Biological five independent experiments were performed. Error bars represent SD. *P < .05 and ***P < .001 (Student’s t-test).

Figure 2.

XPF is methylated at eight arginine sites in cell. (A) Schematic diagram of XPF–ERCC1 showing methylated arginine sites. Pink arrows indicate methylated arginine sites. (B) In vitro methylation assay with CARM1. AAA is CARM1 inactive mutant. MMA is monomethyl arginine and ADMA is asymmetric dimethyl arginine. The bar graph at the bottom shows the relative activity to the methylation activity at R568 of CARM1 WT. (C) Complementation of XP2YO-SV UV sensitivity by mutants in which the XPF arginine site to be methylated is replaced by lysine. XP2YO-SV are fibroblasts derived from XP-F patients. XP2YO-SV cells were transiently transfected with XPF WT and RK-Myc mutants. Cells were irradiated with indicated UV. Biological three independent experiments were performed. Error bars represent SD.

Figure 3.

XPF 6RK with six arginine sites replaced by lysine showed defects in CPD removal, subcellular localization, and UV sensitivity. (A) DOX-inducible XPF WT complemented the CPD removal ability of XP2YO-SV, but XPF 6RK was partially deficient in complementation. VA13 is a normal human-derived fibroblast. Data are shown as the mean ± SD of biological three independent experiments (multiple comparisons: two-way ANOVA, ***P < .001, ****P < .0001). (B) XPF 6RK expressing cells showed decreased total protein content of XPF and decreased chromatin localization. (C) XPF 6RK expressing cells showed increased sensitivity to UV light. Biological three independent experiments were performed. Error bars represent SD. **P < .01, ***P < .001 (Student’s t-test).

Figure 4.

Mapping CARM1-dependent arginine methylation sites onto the human XPF–ERCC1 structure. (A) Overall structure of the auto-inhibited XPF–ERCC1 heterodimer (PDB code 6SXA). Structure is shown as a cartoon using rainbow colors for XPF from the N-terminus to C-terminus highlighting the first RecA domain (blue) and second RecA domain (green) the helical domain (green/cyan), the nuclease domain (yellow), and the HhH₂ domain (red). ERCC1 is colored orange for both its central domain and HhH₂ domain. This figure was prepared using PyMol. (B–F) Close-up views of the structural environment surrounding each of the five putative methylated arginine sites identified as targeted by CARM1. Colors are as described for panel (A).

Figure 5.

XPF R568K with R568 arginine site replaced by lysine showed defects in CPD removal, subcellular localization, UV sensitivity, and accumulation at local UV sites. (A) DOX-inducible XPF WT and R180K complemented the CPD removal ability of XP2YO-SV, but XPF R568K was completely deficient, and R726KR750K were partially deficient in complementation. Data are shown as the mean ± SD of biological three independent experiments (multiple comparisons: two-way ANOVA, *P < .05). (B) XPF R568K expressing cells showed decreased total protein content of XPF and decreased chromatin localization. (C) XPF R568K expressing cells showed increased sensitivity to UV light. Biological three independent experiments were performed. Error bars represent SD. **P < .01, ***P < .001 (Student’s t-test). (D) Accumulation of XPF R568K and ERCC1 at local UV sites was inhibited in R568K expressing HeLa cells. Representative images of immunostaining for pHBO1 and XPF-Myc and pHBO1 and ERCC1 in local UV-irradiated HeLa cells. HeLa cells were irradiated with 50 J m⁻² UV through an 8-μm pore membrane and fixed after 30 min of incubation. Scale bars, 5 μm. Data are shown as the mean (unpaired t-test). (E) In cells expressing XPF R568K to the same extent as WT, XPF R568K-Myc still showed a reduced amount of chromatin bound.

Figure 6.

XPF in CARM1 depleted cells and DOX-induced XPF 6RK and R568K show no change in messenger RNA (mRNA) expression but a marked decrease in protein levels. (A) XPF and ERCC1 proteins in CARM1 depleted cells were significantly reduced. (multiple comparisons: two-way ANOVA, *P < .05, **P< .01). (B) DOX-inducible XPF 6RK and R568K in XP2YO-SV cells had significantly reduced expression of XPF and ERCC1 proteins, although mRNA expression was unchanged. (Multiple comparisons: two-way ANOVA, *P < .05). (C) DOX-inducible XPF R568K in HeLa cells had significantly reduced expression of XPF and ERCC1 proteins, although mRNA expression was unchanged (multiple comparisons: two-way ANOVA, **P< .01). (D) XPF and ERCC1 protein expression in DOX-inducible XPF WT or R568K HeLa cells overexpressing CARM1-WT or AAA in XPF (multiple comparisons: two-way ANOVA, **P< .01). (E) GST pull-down assay using in vitro translated XPF WT and R568K and GST-ERCC1 expressed and purified in E. coli. E1 is ERCC1, RK is XPF R568K-Myc. (F) XPF, which had been subjected to an in vitro methylation assay using affinity-purified CARM, was pulled down using GST-ERCC1. XPF was in vitro translated and affinity purified using MycHis-tag. CARM1 was overexpressed in HEK293 cells and affinity purified using FLAG-tag. GST-ERCC1 was expressed in E. coli and purified using glutathione beads. The bar graph shows the relative binding value against WT (WT−) that was not subjected to the methylation assay. Relative binding of XPF was corrected with GST-ERCC1. (G) Nuclease assay using XPF WT and R568K. XPF-Myc-ERCC1 overexpressed in HEK293 cells was affinity purified with anti-Myc-tag mAb magnetic beads and then incubated with fluorescently labeled oligo DNA substrate for the indicated time. (H) siCt and siCARM1 cells were co-expressed with XPF WT-Myc-ERCC1 and affinity purified, followed by nuclease assays. The assays were incubated for the indicated times. Mock sample was prepared using anti-Myc-tag mAb magnetic beads incubated with mock transfected HEK293 cells.